Intron-mediated enhancement of DIACYLGLYCEROL ACYLTRANSFERASE1 expression in energycane promotes a step change for lipid accumulation in vegetative tissues

Background Metabolic engineering for hyperaccumulation of lipids in vegetative tissues is a novel strategy for enhancing energy density and biofuel production from biomass crops. Energycane is a prime feedstock for this approach due to its high biomass production and resilience under marginal conditions. DIACYLGLYCEROL ACYLTRANSFERASE (DGAT) catalyzes the last and only committed step in the biosynthesis of triacylglycerol (TAG) and can be a rate-limiting enzyme for the production of TAG. Results In this study, we explored the effect of intron-mediated enhancement (IME) on the expression of DGAT1 and resulting accumulation of TAG and total fatty acid (TFA) in leaf and stem tissues of energycane. To maximize lipid accumulation these evaluations were carried out by co-expressing the lipogenic transcription factor WRINKLED1 (WRI1) and the TAG protect factor oleosin (OLE1). Including an intron in the codon-optimized TmDGAT1 elevated the accumulation of its transcript in leaves by seven times on average based on 5 transgenic lines for each construct. Plants with WRI1 (W), DGAT1 with intron (Di), and OLE1 (O) expression (WDiO) accumulated TAG up to a 3.85% of leaf dry weight (DW), a 192-fold increase compared to non-modified energycane (WT) and a 3.8-fold increase compared to the highest accumulation under the intron-less gene combination (WDO). This corresponded to TFA accumulation of up to 8.4% of leaf dry weight, a 2.8-fold or 6.1-fold increase compared to WDO or WT, respectively. Co-expression of WDiO resulted in stem accumulations of TAG up to 1.14% of DW or TFA up to 2.08% of DW that exceeded WT by 57-fold or 12-fold and WDO more than twofold, respectively. Constitutive expression of these lipogenic “push pull and protect” factors correlated with biomass reduction. Conclusions Intron-mediated enhancement (IME) of the expression of DGAT resulted in a step change in lipid accumulation of energycane and confirmed that under our experimental conditions it is rate limiting for lipid accumulation. IME should be applied to other lipogenic factors and metabolic engineering strategies. The findings from this study may be valuable in developing a high biomass feedstock for commercial production of lipids and advanced biofuels. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1186/s13068-023-02393-1.


Background
Production of biofuel from perennial crops will increase energy security by reducing the need for fossil fuels while increasing carbon sequestration for combating climate change [1,2].Therefore, many countries have implemented government policies and subsidies promoting the production and consumption of bioenergy [3].
Energycane, is one of the most promising bioenergy feedstocks.Both energycane and sugarcane are highly polyploid, interspecific hybrids between Saccharum officinarum and Saccharum spontaneum, but energycane has a higher proportion of Saccharum spontaneum in its genome.This contributes to greater biomass yields, elevated resilience under marginal conditions, lower sugar, and higher fiber content in the stems of energycane relative to sugarcane [4][5][6][7].Energycane's superior resilience and biomass accumulation are supported by a more vigorous root system and higher light interception efficiency than sugarcane.The elevated light interception efficiency is associated with a higher tiller density and leaf inclination angles, contributing to the greatest photosynthetic efficiency among crops [8].Photosynthesis, which transforms sunlight into stored chemical energy creates biomass that can be converted to renewable biofuel.Plant oil, consisting mainly of triacylglycerol [TAG; (98%)] and other fatty acids [9,10] has more than twice the energy content of carbohydrates [11,12].Oil is typically produced from seeds of palm and oilseed crops such as soybean (Glycine max), sunflower (Helianthus annus) and canola (Brassica napus) [13].Vegetative plant tissues including leaves, stems and roots are able to produce lipids and TAG in small quantities [14,15].The genes involved in the fatty acid and TAG synthesis, and hydrolysis have been identified [14,16].This enables metabolic engineering of high biomass crops including sugarcane and energycane for the hyperaccumulation of lipids in their vegetative biomass and has the potential to exceed oil yields of traditional oilseed crops [12,[17][18][19][20]. Ectopic expression of individual lipogenic factors led to modest lipid accumulation in vegetative tissues.A step change in accumulation of vegetative lipids has been reported following multigene engineering termed "push-pull-protect strategy" including genes involved in fatty acid biosynthesis ('push'), TAG assembly ('pull'), and suppression of lipid turnover ('protect') [21].This approach optimized the flux of carbon into TAG at multiple metabolic levels, leading to hyperaccumulation of TAG in the vegetative tissues of model [22][23][24][25] and high biomass crops such as ryegrass [26], sugarcane [18,20], sorghum [27], and potato [28].WRI1 is a transcription factor of the family of APETALA2/ethylene-responsive element binding proteins involved in the regulation of fatty acid biosynthesis [29][30][31][32][33]. DGAT 1 catalyzes the only committed step in the biosynthesis of TAG by catalyzing the transfer of an acyl group from acyl-CoA to diacylglycerol [34].DGAT 1 has been classified into three types; two endoplasmic reticulum (ER)-localized forms of DGAT 1 (type-1) and DGAT 2 (type-2) and a cytosolic form of the DGAT family (Type-3) [35,[36][37][38].Overexpression of DGAT 1 alone has been reported to increase the TAG content in the transgenic leaves by 20-fold compared to control [39].Protection of TAG from turnover by lipases has been achieved by coating oil bodies with oleosin [22].OLEO-SIN1 (OLE1) is a structural protein that protects oil bodies from coalescing and reduces lipid breakdown by lipases [40,41].Previous research has shown increased fatty acid content of Arabidopsis vegetative tissues through the engineering of OLE1 (CysOLE1) by the substitution of six cysteine residues for their native equivalents within the N-and C-terminal hydrophilic domains [42].
We earlier reported that in the monocot Lemna japonica, co-expression of WRI1, DGAT 1 and OLE1 had a much larger effect on TAG accumulation than co-expression of WRI1 and OLE1 without DGAT 1, suggesting that the elevated FA synthesis due to ectopic WRI1 expression exceeds the capacity of the native DGAT to transfer FA to TAG [43].Elevating the expression of DGAT is therefore a promising approach for hyperaccumulation of TAG in vegetative tissues.Various optimizations including stronger promoters or enhanced UTR's, codon optimization or the presence of intron(s) have been described to elevate transgene expression in plants [44].Introns can have a surprisingly large effect on gene expression, revealing a gap in our understanding of the molecular basis underlying this phenomenon [45][46][47].
In this study, we found that the addition of an intron from the predicted 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase 1 of Sorghum bicolor into a codon-optimized TmDAGT 1 from nasturtium dramatically increased the gene expression under control of the strong constitutive Panicum virgatum ubiquitin promoter (pPvUbiII).The elevated TmDGAT 1 expression displayed strong synergy in TAG and TFA accumulation following the co-expression of WRI1, DGAT 1, and OLE1 in energycane.

Cis-regulatory elements in the 110-bp intron
A putative GT1 consensus and CAAT box were identified at the position of 52 bp and 93 bp, respectively, in the 110-bp-long intron that we inserted into DGAT 1 and which originated from the predicted 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase 1 of Sorghum bicolor (NCBI accession number: NC_012877.2) (Figure S1).Sequences with similarity to motifs that were earlier reported in IME including YATCN, the reverse complement of NGATY and similar sequence to TTNGAT YTG and its reverse complement CAY ATC NAA were also identified.
Overexpression of WRI1, DGAT 1 and OLE1 altered TAG and TFA production in leaves and stems of transgenic energycane WRI1, DGAT 1, and OLE1 expression was analyzed from leaf and stem tissues by qRT-PCR with primers shown in Table S1.To determine potential intron-mediated regulation of gene expression, we analyzed the expression in five PCR-positive lines for DGAT 1 with intron (DGAT 1 (In)) and compared this with the expression of in five PCR-positive lines for DGAT 1 without intron (DGAT 1 (W)).DGAT 1 (In) elevated expression in leaves by seven times on average of 5 transgenic lines compared to DGAT 1 (W) (Table 1).Expression in stem was two times higher for DGAT 1 (In) compared to DGAT 1 (W) (Table 2).
As shown in Table 1, 8 out of 10 transgenic lines coexpressed all three lipogenic factors with significant differences in transgene expression between the lines.Two of the transgenic lines (L4; L10, Table 1) expressed DGAT 1 and OLE1 but not WRI1.Among the three transgenes, WRI1 expression showed the lowest variation ranging from 0.01 to 0.05 relative to GAPDH and averaged 0.02 across the five DGAT 1 (W) or DGAT 1 (In) lines (Table 1).In WDO combination, TAG accumulated up to 1.01% of leaf DW and 0.22% of stem DW, which was 50-fold and tenfold of that of WT, respectively (Tables 1,  2).The highest TAG accumulation was observed in WDiO combination with up to 3.85% of leaf DW and 1.14% of stem DW, which was 192-fold and 56-fold of that of WT, respectively (Tables 1, 2).Similarly, TFA was the highest in WDiO combination with up to 8.39% of leaf DW and 2.08% of stem DW (Tables 1, 2).In addition, Pearson's correlation was evaluated for TAG accumulation and lipid gene expression levels in transgenic plants.As shown in Table S2, TAG accumulation exhibited the highest positive correlation with DGAT 1(W) (r = 0.97) followed by OLE1 (W) (r = 0.89), DGAT 1 (In) (r = 0.88), and WRI1 expression (r = 0.51).

Accumulation of lipid droplets in leaves of transgenic energycane with hyperaccumulation of TAG in comparison to WT controls
For the purpose of visualizing lipid droplets, leaves from both the transgenic line L13 and its corresponding WT were subjected to boron-dipyrromethene (BODIPY) staining, followed by imaging using confocal microscopy.Notably, the transgenic line L13, with the overexpression of DGAT 1 (In), demonstrated a substantial increase in the abundance of BODIPY-stained lipid droplets in comparison to its WT counterpart (Fig. 3).Intriguingly, within the transgenic line L13, the  lipid droplets, encompassing both regular lipid droplets (LD) and guard cell lipid droplets (GCL), exhibited a widespread distribution.The most prominent droplet size, around 8 μm in diameter, was predominantly situated within the stomatal guard cells.In contrast, the wild-type leaves exclusively exhibited lipid droplets within stomatal guard cells, typically measuring less than 5 μm in diameter.These findings correlate well with the triacylglycerol (TAG) data depicted in Table 1.

Impact of hyperaccumulation of TAG on plant growth
To determine the effect of hyperaccumulation of TAG on plant growth, biomass related traits, including plant height, stem diameter, and tiller number were recorded.Hyperaccumulation of TAG corresponded to a reduction in plant height, tiller number, and stem diameter Table 3. Line 13 with the highest TAG accumulation had the greatest reduction in plant height, number of tillers and stem diameter with 31%, 81% and 50%, respectively.In addition, plant height, tiller number and stem diameter correlated negatively with TAG content (r = − 0.71, − 0.77 and − 0.50, respectively) (Additional file 2: Table S3).

Discussion
Energycane is among the most promising perennial feedstocks to contribute to the emerging bioeconomy due to its high biomass production, yield resilience and ease of biocontainment [4][5][6]48].Metabolic engineering for hyperaccumulation of lipids in vegetative tissues has been proposed for enhancing energy density and biofuel production from high biomass crops like energycane [12,17,20,21].Metabolic engineering of energycane is still in its infancy due to the recalcitrance of this crop to tissue culture and genetic transformation [17].Here we report a step change in hyperaccumulation of triacylglycerol (TAG) and total fatty acid (TFA) in energycane using intron-mediated enhancement (IME) of the expression of DGAT 1. Line-to-line variation in transgene expression in different transgenic events has been reported due to epigenetic changes [49,50], position of transgene insertion in the genome [51,52] and copy number of the transgene [53][54][55].Therefore, several transgenic events need to be analyzed per recombinant DNA construct to support conclusions on the impact of specific approach/construct.In this study, five lines were evaluated for each, TmDGAT 1 with and without intron.Including an intron in TmD-GAT 1 (TmDGAT 1 (In)) increased its expression sevenfold compared to the intron-less variant.Combination of TmDGAT 1 (In) expression with SbWRI1 and SiOLE1 expression elevated TAG accumulation 3.8-fold compared to the intron-less combination.Our findings also confirm that DGAT 1 is a rate-limiting enzyme for TAG accumulation in vegetative tissues of energycane.
The potential of introns to elevate transgene expression by up to 100-fold, termed intron-mediated enhancement (IME) has been described more than 30 years ago [45,[56][57][58][59][60][61].IME can strongly stimulate mRNA accumulation far downstream of the transcription start site, even in the absence of a promoter, revealing gaps in our understanding of the molecular mechanisms governing gene expression [62].Identification of introns that confer IME requires systematic evaluation in transgenic cells since not all introns lead to transcription enhancement [62].Introns can differ in the presence of enhancers [63,64] or transcription factor binding sites [65].In addition, splicing can influence transcription by affecting the phosphorylation state of RNA polymerase II [66].Introndependent gene looping allows physical interaction of the promoter and the terminator regions accelerating mRNA processing events including capping and polyadenylation [67][68][69].Disposition of the exon junction complex proteins also enhances mRNA stability, export and translation [70][71][72].
The impact of introns on transgene expression is influenced by a variety of factors including position of introns [73][74][75], length of exons [76], specific intronic sequence elements [63,74,77], splicing efficiency [74,76] and their ability to promote the formation of a multi-looped gene architecture [47,74].Therefore, not all introns lead to elevated gene expression [78].Recent studies have identified specific cis-elements important for IME expression [46,[79][80][81][82][83].The intron that was used in this study is constitutively spliced and contains the NGATY core of the longer TTNGAT YTG motif and two additional motifs that are similar to TTNGAT YTG [46,83].These cis-elements could play a role in IME, in addition to reduced exon length and supporting transcript processing and nuclear export with the constitutive splicing mechanism.The online transgene design tool "Intronserter": https:// bibis erv.cebit ec.uni-biele feld.de/ intro nsert er supports systematic intron insertion into transgenes for reduction of exon length, along with codon optimization and removal of undesired sequence elements.The program is fully customizable for individual eukaryotic target organisms and should support the systematic design of transgenes for IME in support of metabolic engineering strategies [84].
In the current study, by co-expression of WRI1, DGAT 1 (In), and OLE1, we showed an average TAG hyperaccumulation of 3.85% of leaf DW.This TAG content was 192-fold higher than in WT energycane and 2.5-fold higher than reported recently for energycane [17].The hyperaccumulation of total fatty acid was also elevated to 8.4% of leaf DW which is almost 56 times that of WT and 1.7-fold than in the first report of metabolic engineering of energycane [17].High level of TAG and TFA accumulation in vegetative tissues of plants typically requires co-expression of several lipogenic factors involved in upregulation of fatty acid biosynthesis, TAG biosynthesis and its protection from degradation [18,20,23,27].Constitutive co-expression of WRI1, TmDGAT 1 (In), and OLE1 in this study led to 2.5-fold higher TAG and 1.9fold higher TFA accumulation in energycane than using ZmDGAT 1-2 without intron and OLE1 in combination with RNAi co-suppression of the TAG lipase sdp1 and tgd1, a lipid transporter from the endoplasmic reticulum (ER) to chloroplasts [17].We expect the optimization of expression cassettes using IME for all lipogenic factors, codon optimization, stacking of additional lipid genes in combination with gene editing approaches [85][86][87][88][89] will further elevate oil accumulation in energycane.
Reported TFA contents following ectopic co-expression of WRI1, DGAT 1, and OLE1, in grain sorghum [27] (up to 9.7% of DW) or tobacco [24] (up to 17.7% of DW) exceed the levels of TFA reported here for energycane (8.4% of leaf DW) by 1.16-to 2.11-fold, respectively.However, biomass yield for energycane are typically in the range of 41 to 49 t ha −1 in the Southeastern USA [90,91] which is 3-to 12-fold higher than typical biomass yields from grain sorghum (4 to 12 t ha −1 ) [92,93] and 3-to 9-fold higher than typical tobacco yields (5.5 to 13 t ha −1 ) [94,95].This should lead to higher or similar oil yields from the energycane described here compared to the alternative crops.Biodiesel derived from lipid accumulating energycane is an attractive approach to minimize the use of food and feed crops in the production of renewable transportation fuels.Additional feedstocks and engineering strategies will be needed to establish a robust pipeline for production of sustainable aviation fuel [96].
Coordinating the expression of WRI1 with stem specific or developmentally regulated promoters active late in plant development [97,98] or inducible promoters [43,99] may mitigate the negative impact on biomass accumulation due to constitutive expression of WRI1 and other lipogenic factors described in this and several other studies [18,25,27].
The overall TAG and TFA compositions were significantly altered by the co-expression of the lipogenic factors in energycane.Most notable was the increase in oleic acid (18:1) and linoleic acid (18:2) in both, TAG and TFA of transgenic leaves and stems at expense of saturated fatty acids.Similar findings were earlier reported for sugarcane by Parajuli et al. [18].A significant reduction of alpha linoleic acid (18:3) was only observed in the TFA of leaves.Interestingly in stem tissues oleic acid (18:1) linoleic acid (18:2) and alfa-linolenic acids (18:3) were significantly elevated in both TAG and TFA which differs from the observation in sugarcane using different lipogenic factors [18].The elevated amount of unsaturated FA in vegetative tissues of energycane should lead to improved fuel stability and cold flow [100].
The presence of leaf GCL droplets as seen in guard cells of WT have been reported for more than three decades [101], where they serve as an energy source to power stomatal opening [102].Subsequent work on Arabidopsis and the lycophyte Selaginella provide evidence that TAG breakdown is an evolutionarily conserved mechanism in light-induced stomatal opening [103].In the oil accumulating L13, constitutive expression of WRI1, DGAT 1 and OLE1 resulted in a substantial boost in GCL abundance, and LDs were generally visible in surrounding tissue, that were absent in WT controls.The pattern of widespread LD accumulation in transgenic energycane leaves is consistent with reports from transgenic sugarcane [18], tobacco [21], and the duckweed Lemna japonica [43].

Conclusions
Energycane is a favored feedstock for the production of bioproducts and advanced renewable fuels because of its high biomass production and resilience under marginal conditions.In this study, we demonstrated that DGAT 1 is the principal rate-limiting step for TAG and TFA accumulation in energycane leaves.IME of DGAT 1 caused a step change in its expression and its co-expression with WRI1 and OLE1 elevated TAG or TFA content by 192-or 56-fold relative to levels found in the leaves of unmodified energycane, respectively.These results lay the foundation for the commercial production of biodiesel and other FA derivatives and reveal opportunities for further optimizations of high biomass energycane.

Design and synthesis of the optimized DGAT 1 gene
For codon optimization, the native DGAT 1 gene from Tropaeolum majus (Gene bank AY084052) was optimized by replacing the preferred codons with synonymous codons of Sorghum bicolor, a close relative of sugarcane, without changing the encoded amino acids (Additional file 1: Figure S1).Two optimized DGAT 1, including DGAT 1(W) without intron and DGAT 1 (In) with the 110-bp-long intron from the predicted 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase 1 of Sorghum bicolor (NCBI Accession number: NC_012877.2),were synthesized by GenScript (Piscataway NJ, USA).The specific intron was selected based on its compact size and the presence of cis-regulatory elements that may impact IME (Additional file 1: Figure S1).

Construction of the multigene expression vectors
Two different optimized DGAT 1 genes obtained from Tropaeolum majus, including DGAT 1(W) without intron and DGAT 1 (In) with a 110-bp-long intron were compared in this study.Using the Golden Gate cloning strategy [104,105], three multigene expression vectors were created (Fig. 4).The first vector, shown in Fig. 4A, contains WRI1 from Sorghum bicolor controlled by the Brachypodium distachyon elongation factor 1-α promoter (pBdEF1α) and Panicum virgatum ubiquitin (tPvUbiII) terminator, and the selectable marker gene neomycin phosphotransferase (nptII) driven by the maize ubiquitin promoter (pZmUbi) and Sorghum bicolor heat shock protein terminator (tSbHSP; Fig. 4A).Vectors shown in Fig. 4B and C contained expression cassettes for two genes: OLE1, obtained from Sesamum indicum, was under the control of the Brachypodium distachyon ubiquitin promoter (pBdUbi10) and Nicotiana benthamiana actin 3' UTR terminator (tNbACT3), and DGAT 1 from Tropaeolum majus, with or without intron was cloned under the control of the Panicum virgatum ubiquitin promoter (pPvUbiII) and Nicotiana tabacum extensin terminator (tNtEU; Fig. 4B and C).

Tissue culture and genetic transformation of energycane
For this study, the energy cane genotype UFCP 84-1047 (WT) was chosen for its superior biomass and resilience [106].Immature leaf whorl cross-sections (Fig. 5A) were cultured to induce embryogenic callus (Fig. 5B) as described by Fouad et al., [107].The media for callus induction, transformation, selection, and shoot and root regeneration were prepared according to Fouad et al. [107].

Cis-regulatory elements in the 110-bp intron
PLACE and PlantCARE software were used to evaluate the presence of potential cis-regulatory elements in the 110-bp-long intron from the predicted 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase 1 of Sorghum bicolor (NCBI accession number: NC_012877.2).We also searched for sequence motifs that were earlier reported for their contribution to IME including sequence TTNGAT YTG and its reverse complement CAY ATC NAA as well as NGATY (or its reverse complement YATCN) [76].

PCR analysis of genomic DNA extracts
Genomic DNA was extracted from leaves using the cetyl trimethyl ammonium bromide (CTAB) method [109].100 ng of genomic DNA was used as a template in a 20-μL PCR reaction.PCR amplification was conducted using Hot Start Taq DNA Polymerase (NEB, Ipswich, Massachusetts, USA) for WRI1, nptII, OLE1, and DGAT 1 individually following the manufacturer's protocol.Primers for each target gene are listed in Additional file 2: Table S1.Five transgenic lines in each combination were confirmed by PCR for WRI1, DGAT 1, OLE1, and nptII (Additional file 1: Figure S2).

Greenhouse propagation of transgenic, lipid accumulating energycane
In the greenhouse, transgenic energycane (Fig. 5E) was propagated by nodal stem cuttings to obtain biological replicates and three plants per transgenic line and non-transgenic plants were each planted in a pot with a 15 cm diameter containing potting mix (Jolly Gardener C/G, Oldcastle Lawn and Garden, Atlanta GA).Plants were irrigated and fertilized daily with a drip fertigation system.In the greenhouse, the temperature was controlled by evaporation cooling to 25-29 °C during the day and 20-24 °C during the night using natural photoperiod with a maximum daily light intensity of 1000 to 1500 μmol m −2 s −1 .

Sampling of leaves and stem for quantitative real-time PCR (qRT-PCR), TAG and FA analysis
Leaves from 15 transgenic lines were sliced in half at the midrib; one half was used for qRT-PCR analysis, while the other half was used for TAG and FA analyses.Stem internodes from the three highest TAG lines were also sampled and divided into two halves before grinding in a Retsch cryo bead mill (Verder Scientific, Haan, Germany).For qRT-PCR analysis, samples were flash-frozen in liquid nitrogen and maintained at − 80 °C for later RNA extraction.For TAG and FA analyses, each sample consisted of at least 100 mg of leaf or stem fresh weight and was freeze-dried for 72 h in a lyophilizer (Labconco, MO, USA).For TAG and FA analyses samples were shipped to Brookhaven National Laboratory on dry ice.

qRT-PCR analysis for transgene expression
TRIzol reagent was used to extract total RNA from leaf/ stem samples (Life Technologies, Thermo Fisher Scientific, Waltham, MA, USA).A High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) was used to synthesize cDNA from approximately 1.0 µg of total RNA from each sample.DGAT 1, OLE1, and WRI1 expressions were measured using primers shown in Additional file 2: Table S1.The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene was used as a housekeeping gene for transcript normalization [110].qRT-PCR was performed in a Real-Time PCR Detection System (Biorad, Hercules, CA, USA) using SsoAdvanced Universal SYBR green supermix (Biorad, Hercules, CA, USA).The formula 2 {Ct (GAPDH) -Ct (transgene)} , was used to determine the relative transcription levels of WRI1, DGAT 1, and OLE1.

Analysis of TAG and total FA composition
TAG and FA analyses were conducted according to Zale et al., 2016 [20].In brief, 10.0 mg of lyophilized leaf/stem tissue was extracted with 700 µL of extraction solution containing 2:1:0.1 by volume methanol:chloroform:formic acid.After 1 h of mixing on a vortex mixer and then standing at RT overnight, extract total lipids were as the lower phase by adding 350 μL of 1 M KCl and 0.2 M H 3 PO 4 followed by briefly vortexing and centrifuging at 2000 g for 10 min Total lipids were separated by thin layer chromatography (TLC) using a mixture of hexane, diethyl ether and acetic acid (70:30:1, by volume) as mobile phase and TLC Silica gel 60 (EMD Millipore, Cat.No. 1056260001).By spraying 0.05% primuline (in 80% by volume of acetone), lipids were visualized under UV light.TAG fractions were scraped from the TLC plate after identifying by comparing mobility with that of authentic TAG standards.After being incubated in 1.0 mL of boron trichloride-methanol for 40 min at 80-85 °C, TAG fractions were transmethylated into FA methyl esters (FAMEs).FAMES were extracted into hexane and dried under a nitrogen stream before being dissolved in 100 μl hexane and analyzed by GC-MS with an Agilent Technologies (Santa Clara, CA, USA), 7890A GC System equipped with an Agilent 60 m DB23 capillary column (ID 250-μm), and a 5975C mass selective detector in full mass scan mode.The oven temperature was ramped from 100 °C to 240 °C at a rate of 15 °C min −1 and held at 240 °C for 3 min with a flow rate of 1.2 ml min −1 .Following incubation in 1.0 mL of boron trichloride-methanol, total lipid extracts were immediately transmethylated into FAMES for total FA analysis.Using gas chromatography-mass spectrometry, FAMES was dissolved in 100 µL hexane and measured.An internal standard of 5.0 g of C17:0 was employed.

Visualization of lipid droplets with confocal microscopy
For the visualization of lipid droplets, the third leaf from the non-necrotic tip region of both the transgenic line L13 and its corresponding WT was carefully excised.Subsequently, these leaf samples underwent fixation using a FAA buffer (4% formalin, 5% glacial acetic acid, 50% ethanol, 41% water, V/V) under vacuum conditions for a duration of 1 h.Following fixation, the samples were stained with a solution containing 12 μg/mL of BODIPY 493/503 (Invitrogen, Eugene, OR, USA) and 0.1% Triton X-100, employing vacuum assistance for 20 min.The imaging process was carried out using a Leica TCS SP5 laser scanning confocal microscope, employing an excitation wavelength of 496 nm for BODIPY and capturing emissions within the range of 505-583 nm.Additionally, chlorophyll autofluorescence was captured within the wavelength range of 661-800 nm.

Statistical analysis
The TAG content, total FA content, biomass, and qRT-PCR analysis data were expressed as means ± standard error (SE).To conduct statistical analysis of the means, one-way ANOVA was completed using SPSS version 20.0 (Armonk, NY, USA).A value of p ≤ 0.05 was considered statistically significant.Pearson's correlation coefficient was also calculated using SPSS.For each statistical analysis, three independent biological replications were used.Award number DE-SC0018254 supported vector construction, generation of transgenic plants and their molecular analysis.Award Number DE-SC0018420 supported the lipid analysis of transgenic plants.The confocal microscopic study at the Center for Functional Nanomaterials at Brookhaven National Laboratory was supported under Office of Basic Energy Sciences of DOE with contract no.DE-SC0012704 (to CFN).The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.This work was also supported by the USDA National Institute of Food and Agriculture, Hatch project 1020425.

Fig. 1
Fig.1TAG FA composition and total FA composition in leaves of different transgenic energycane lines (L11; L12; L13) compared with non-modified energycane (WT).A TAG composition in leaves of different transgenic lines (L11; L12; L13) and non-modified energycane (WT).B TFA composition in leaves of different transgenic lines (L11; L12; L13) and WT.Vertical bars are the means ± standard error (n = 3).Values with different letters are significantly different at p ≤ 0.05 according to a one-way ANOVA test and the Duncan's multiple range test (MRT).TAG: triacylglycerol.FA: fatty acid.WT: non-modified energycane (WT)

Fig. 2
Fig.2TAG FA composition and total FA composition in stems of different transgenic lines (L11; L12; L13) compared with non-modified energycane (WT).A TAG composition in stems of different transgenic lines (L11; L12; L13) and non-modified energycane (WT).B TFA compositions in stems of different transgenic lines (L11; L12; L13) and WT.Vertical bars are the means ± standard error (n = 3).Values with different letters are significantly different at p ≤ 0.05 according to a one-way ANOVA test and the Duncan's multiple range test (MRT).TAG triacylglycerol.FA fatty acid.WT non-modified energycane (wild-type)

Fig. 3
Fig. 3 Lipid droplets within the leaf tissue of energecane.The upper panel showcases fluorescence-stained lipid droplets within guard cells (GCL) using BODIPY 493/503, along with the autofluorescence emanating from chloroplasts in the wild-type (WT) strain.The lower panel displays both the lipid droplets within guard cells and those in regular cells (LD), alongside the chloroplast autofluorescence found in leaf tissue of the transgenic line L13.The lipid droplets within guard cells are indicated by white arrows, while a selection of typical lipid droplets are indicated by blue arrows.Both the lipid droplets found within the guard cells (GCL) and those present outside of the guard cells (LD) are rendered in shades of green, while autofluorescence originating from chlorophyll is is seen in red.Scale bar = 50 μm

Fig. 5
Fig. 5 Generation of transgenic energy cane.A Leaf whorl cross-sections were used as explants.B Callus induced from leaf whorl cross-section was used as the target for biolistic gene transfer.C Calli regenerating on medium with geneticin for selection of nptII expression.D Regenerated plants before transferring to soil.E Transgenic energy cane lines growing in greenhouse.Bar = 5 mm

Table 1
Summary of TAG and TFA content and transgene expression in leaves of energycane TAG values and gene expression are shown for each line representing leaf extracts from three replications.The leaf samples used are the first dewlap leaf.Values are means ± SE (n = 3).The expression of transgenes is shown relative to GAPDH.WT indicates the non-transgenic plants.WDO indicates the transgenic lines expressing WRI1, DGAT 1 without intron, OLE1 and nptII.WDiO indicates the transgenic lines expressing WRI1, DGAT 1 with intron, OLE1 and nptII.Bold values indicate the mean of all WDO or all WDiO lines, respectively.Values in the same column with different lower case letters are significantly different at p ≤ 0.05

Table 2
Summary of TAG and TFA content and transgene expression in stems of energycane TAG values and gene expression are shown for each line representing stem extracts from three replications.Values shown for each line represent the mean of mature, mid-mature and immature stem samples from three replications.Values are means ± SE (n = 3).The expression of transgenes is shown relative to GAPDH.WT indicates the non-transgenic plants.WDO indicates the transgenic lines expressing WRI1, DGAT 1 without intron, OLE1 and nptII.WDiO indicates the transgenic lines expressing WRI1, DGAT 1 with intron, OLE1 and nptII.Bold values indicate the mean of all WDO or all WDiO lines, respectively.Values in the same column with different lower case letters are significantly different at p ≤ 0.05 Construct Lines TAG (% of DW) TFA (% of DW) Transgene expressions normalized to GAPDH